Hacker News new | ask | show | jobs
by xnx 555 days ago
This would be super useful as a training set for automatically identifying ads in podcasts that contain non-ad content.
1 comments

If you produced the podcast this would be trivial since you’d know the time stamps for the ad holes.
Yes, though I don't think you know how long the ad block will be. I've seen some podcast MP3 files that have the ad spots indicated in the ID3 tag, but that's not super useful because it only gives the original start points.
It'll penalize advertiser provided scripts more.